Available online at www.sciencedirect.

com

ScienceDirect
Energy Procedia 75 (2015) 118 – 123

The 7th International Conference on Applied Energy – ICAE2015

Ammonia, Methane and Hydrogen for Gas Turbines

Valera-Medina Aa*, Morris Sa, Runyon Ja, Pugh DGa, Marsh Ra,
Beasley Pb, Hughes Tb
a
Cardiff University, Queen’s Building, Cardiff CF24 3AA, United Kingdom
b
Siemens, 23-38 Hythe Bridge St, Oxford OX1 2EP, United Kingdom

Abstract

Ammonia has been identified as a sustainable fuel for transport and power applications. Similar to hydrogen,
ammonia is a synthetic product that can be obtained either from fossil fuels, biomass or other renewable sources.
Since the 1960’s, considerable research has taken place to develop systems capable of burning the material in gas
turbines. However, it is not until recently, that interest in ammonia has regained some momentum in the energy
agenda as it is a carbon free carrier and offers an energy density higher than compressed hydrogen. . Therefore, this
work examines combustion stability and emissions from gaseous ammonia blended with methane or hydrogen in gas
turbines. Experiments were carried out in a High Pressure Combustion Rig under atmospheric conditions employing
a bespoke generic swirl burner. OH* Chemiluminescense was used for all trials to determine reactivity of the radical.
Emissions were measured and correlated to equilibrium calculations using GASEQ. Results show that efficient
combustion can be achieved with high power but at very narrow equivalence ratios using both hydrogen and methane
blends. Moreover, low concentrations of OH radicals are observed at high hydrogen content, probably as a
consequence of the high NH2 production.
© 2015
© 2015TheTheAuthors. Published
Authors. by Elsevier
Published Ltd. This
by Elsevier Ltd.is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Selection and/or peer-review under responsibility of ICAE
Peer-review under responsibility of Applied Energy Innovation Institute
Keywords: Ammonia; Hydrogen; Stability; Gas Turbines.

1. Introduction
Fuel flexibility is one of the major areas of investigation for gas turbines research. Sources that can
produce minimal or zero CO2 , are currently being investigated by various groups, with the goal of
identifying; clean sources with high power output and low emissions to cope with energy demand, whilst
reducing climate change. Ammonia has been considered as a potential candidate for many years and is
more cost effective than pure hydrogen. Ahlgren [1] performed a study where he compared the fuel power
density F (a unit that takes into consideration the power in the fuel and the energy required to transport it
in pipelines) for different fuels. He concluded that ammonia and methanol are the best renewable fuels
options if compared with pure hydrogen. However, there is still very little literature describing ammonia
in complex flows such as those used in gas turbines.
Karabeyoglu et al. [2] present some important issues on the use of ammonia, arguing that toxicity and
low energy density are the key issues. They also report the development of technologies to burn NH3 in
gas turbines with the objective of minimizing NOx and NH3 emissions in the flue gases. However, most

* Corresponding author. Tel.: +44-2920-875948.

E-mail address: valeramedinaa1@cardiff.ac.uk

1876-6102 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of Applied Energy Innovation Institute
doi:10.1016/j.egypro.2015.07.205
 A. Valera-Medina et al. / Energy Procedia 75 (2015) 118 – 123 119

of these technologies are still in their early stages. Grcar et al. [3] combined experimental and modelling
investigation of ammonia chemistry in a hot combustion environment that was below adiabatic flame
temperatures. The research concluded that the final products of NH3 oxidation remained sensitive to
mixing even at temperatures below those of self-sustaining flames. At low temperatures, NH3oxidation
occurred in a premixed reaction zone, but at sufficiently high temperatures, a non-premixed reaction zone
that produces significantly less NO than the equivalent premixed system might develop.
SPG Advanced Propulsion and Energy [4] gives a coherent history of the use of ammonia in turbojet
systems and Solar used a 250HP T-350 single can burner to run experiments in the 1960’s. Performance
of the engine using ammonia was comparable to that with JP-4, where ammonia at 2.35 times the quantity
of hydrocarbon fuels provided cooler turbine inlet temperatures.
Pratt [5] and Verkamp et al. [6] have shown that the partial cracking of ammonia into hydrogen and
NH is also beneficial, as this blend approximates the properties and burning velocities of methane and
propane, thus providing better efficiencies, higher stability, lower NOx and high levels of power.
Iowa State University [7] have designed a system capable of producing 40kW of power, with nozzles,
swirl stabilisers and a self-sustained heat exchanger similar to those employed in gas turbines. The group
found that the use of a flame holder increased the efficiency of the combustion process, with a clear
reduction of NOx emissions due to the recirculation of the molecule and its interaction with the amines in
the dissociated molecules.
Although there have been developments in this area, further tests are still required to produce a stable,
high power gas turbine capable of producing energy at an industrial scale using ammonia with hydrogen
and/or methane. Therefore, this work is focused on looking at different blends of fuel in a generic swirl
burner relevant to industrial systems. Numerical and experimental analyses were performed in order to
provide enough data for future works for close-to-industrial prototypes.
2. Setup
2.1. Experimental trials
Experiments were performed at the Gas Turbine Research Centre (GTRC). The GTRC is a facility
which consists of two major combustion rigs each designed for detailed investigation of combustion and
emissions characteristics at representative gas turbine conditions. One of these rigs, the High Pressure
Combustion Rig (HPCR) can perform at pressures up to 16bara and deliver air at 5 kg/s at temperatures
reaching 900K. For this project a High Pressure Optical Chamber (HPOC) was fitted with a generic lean
pre-mixed swirl burner, Fig. 1. Quartz windows allocated at both sides of the combustion chamber and in
the exhaust duct allow the burner exit to be viewed from both axial and radial positions. A tangential
swirl burner, relevant to Siemens devices, was designed based on previous experience and numerical
models from past projects.
The GTRC has a unique gas mixing facility that can blend up to five gas streams by mass, in real time,
enabling the effect of variations in fuel compositions on the combustion process to be investigated.
Coriolis mass flow meters are used on the fuel and air lines allowing the accurate control of equivalence
ratio, Re numbers and ammonia/methane/hydrogen concentrations.
OH* chemiluminescense measurements were taken on atmospheric CH4, NH3 -H2 and CH4 -NH3
swirl flames. The chemiluminescense intensity of the OH radical is low and requires the use of an image
intensifier to capture the light emitted in the UV range. A Dantec Dynamics Hi Sense Mk II CCD camera
with a 1.3 megapixel resolution was coupled to a Hamamatsu C9546-03L image intensifier was used. A
specialty 78 mm focal length lens (F-stop = f/3.8) capable of capturing light in the UV wavelength range
was installed on the image intensifier along with a narrow band pass filter centered at 307 nm (FWHM =
10 nm).
120 A. Valera-Medina et al. / Energy Procedia 75 (2015) 118 – 123

Fig. 1. Drawing of Generic Swirl Burner.

Two hundred OH* chemiluminescense images were taken at 10 Hz at each test condition using
Dantec’s DynamicStudio software, while the image intensifier gain was controlled via remote control.
All images were taken through the top window of the HPOC at a 90° angle to the direction of flow.
Two pieces of equipment were used to characterise the produced emissions. The first is an integrated
system developed by Signal Instruments comprising several analyzers: a Flame Ionisation Detector (FID)
is employed within a Signal 3000HM to detect total Hydrocarbons (THCs); a heated vacuum
chemiluminescence analyser (Signal 4000VM, spanned to 900ppm C3H8 equivalent) to quantify NOx
concentrations, calibrated to 37.1 ppm NO and 1.9 ppm NO 2 ; a multi-gas analyser (Signal MGA),
containing an infrared cell for measurement of CO (calibrated for 0-900 ppm) and CO2 (0-9%), in
addition to a paramagnetic O2 sensor (up to 22.5%). A new system has been purchased for quantifying
trace concentrations of NH3 in the flue gas, and forms a subsystem to the 4000VM NOx analyser.
However, this will not be ready until mid-autumn. The second system is a standalone Rosemount NGA
2000 multi-gas analyser, and provides secondary readings of CO, CO 2 and O2 all calibrated to the same
concentrations previously stated.
2.2. Numerical studies
Equilibrium calculations were obtained in order to evaluate all different blends using the software
GASEQ. GASEQ is software written in Visual Basic 3 for combustion equilibrium calculations of
different types of problem. The fundamentals of the program are based on the complex, balanced
chemical calculations defined by Sanford Gordon and Bonnied J. McBridge for NASA [8].
3. Experiments and Discussion
3.1. Test parameters
A reference flame using methane was first obtained and characterized. In order to increase the
ammonia concentration, trials were conducted using air and methane, the latter being slowly reduced
whilst the ammonia was increased. Initially an attempt was made to reach 100% ammonia combustion;
however, the flame blew off before reaching stable conditions. Concentrations are represented here by
volume and similarly, 90%NH3 -10%CH4 concentrations did not provide any results. It was at 80%NH3 -
20%CH4 and 66%NH3 -33%CH4 concentrations that quasi-stable flames were obtained.
Based on the observations from a short constant volume bomb study conducted at the GTRC,
investigating the burning rate of NH3 -H2 blends it was found that a hydrogen concentration close to 50%
would produce similar laminar burning flame velocities to that of methane. Therefore, a 50%NH3 -50%H2
blend was also characterized. All test conditions are presented in Table 1.
 A. Valera-Medina et al. / Energy Procedia 75 (2015) 118 – 123 121

Table 1. Parameters defined for the text matrix

MASS NH3 MASS CH4 /H2 Base POWER
[g/s] [g/s] MASS O2 [g/s] MASS AIR [g/s] [KW] STOICHIOMETRY
Ammonia 80%, Methane 20%
0.800 0.236 2.092 9.472 30.977 0.900
0.800 0.236 1.883 8.525 30.977 1.000
0.800 0.236 1.552 7.026 30.977 1.213
Ammonia 66%, Methane 33%
0.584 0.340 2.592 11.733 31.840 0.740
0.580 0.340 2.096 9.570 31.750 0.910
0.585 0.350 1.733 7.843 32.413 1.108
0.578 0.340 1.666 7.540 31.705 1.143
0.582 0.350 1.561 7.065 32.345 1.220
0.578 0.340 1.491 6.749 31.705 1.277
Ammonia 50%, Hydrogen 50%
0.800 0.094 4.432 19.021 31.349 0.425
0.800 0.094 4.218 18.101 31.349 0.446
0.800 0.094 3.998 17.160 31.349 0.471
0.800 0.094 3.504 15.038 31.349 0.537
0.800 0.094 3.262 14.001 31.349 0.577

3.2. Results
It was observed from experimental trials with the generic burner that the flame would present different
stabilization patterns based on equivalence ratio. As was predicted the methane flame presented a very
stable profile with coherent emissions, showing an increase of NOx and CO at higher equivalence ratios,
with peak combustor exhaust temperatures close to an equivalence ratio of 1. In the case of ammonia and
methane, it was observed that at lean conditions with an equivalence ratio ~ 0.7, the methane appeared to
be burning efficiently at the root of the flame near the exit of the burner with ammonia burning less
efficiently downstream.Evidence to support this hypothesis was found in the gas analysis which measured
high levels of NOx, but low levels of CO and THCs for this condition. It is believed that the unburnt
ammonia was being converted and measured as NOx in the NOx analyser. However, at equivalence ratios
greater than 1.0, the flame started to show a shift in behaviour and shape, probably as a consequence of a
combined effect between the reduction of oxygen and greater dilution of the methane in the total
concentration.
This was evident at an equivalence ratio of 1.05 at which point the flame appears to be incapable of
burning both molecules efficiently. NOx emissions start to reduce, with a considerable decrease in O2 ,
but with increasing CO in the exhaust gases. At this point, the flame starts to show significant signs of
thermoacoustics instabilities most likely caused by the slow, uneven reaction of the blend. A stable
orange flame was observed at an equivalence ratio of 1.1 probably as a consequence of a high content of
NH2 radicals [9], Fig. 2. As the air flowrate was decreased, the flame started to become more unstable
with NOx emissions dropping considerably with a reduction in combustion exhaust temperatures of 60°C
when compared to the hottest conditions measured close to an equivalence ratio of 1.0. Unburnt
hydrocarbons increased substantially in the range of thousands of ppm with the flame showing
considerable corrugation at this condition. Due to the reduced bulk volumetric flow rate through the
burner the system started to experience flashback. This phenomenon was particularly interesting, since
flashback would occur and then the flame would re-position itself at the outlet of the nozzle after a few
seconds. This is probably caused due to the different burning rates of ammonia and methane.
122 A. Valera-Medina et al. / Energy Procedia 75 (2015) 118 – 123

Fig. 2. Flame with an ammonia-methane blend. Equivalence ratio 1.14.

The hydrogen and ammonia blend showed similar trends, but with a narrower band of equivalence
ratios where stability occurred. Unfortunately, it proved impossible to operate the burner close to an
equivalence ratio of 1.0 due to the relatively low bulk volumetric flowrates used and high H2 content.
The flame shape and behavior at the lowest equivalence ratio was very erratic, close to blowoff.
Increasing the equivalence ratio was observed to improve flame stability leading to a regime with
relatively low emissions (NOx 220-260ppm). As the equivalence ratio was increased further, combustion
remained close to the burner exit nozzle with a very intense hydrogen flame and what appeared to be
reasonable NH3 combustion downstream. Emissions showed the highest NOx values (~4500ppm) at
0.577 equivalence ratio before flashing back into the burner.
Chemiluminescense showed clear OH reactions in the methane base flame. When compared to the NH 3
-CH4 case, the results denoted much less OH, a phenomenon that was thought to be a consequence of the
lower reaction of the molecules and increase of the CH radicals in the process. However, it was surprising
that the NH3-H2 process showed a reduced, more uniform OH production when compared to pure
methane. It is believed that the cause of this behavior is the increase of NH2 reactions that not only
produce an orange flame and block the OH emission, but will also consume NOx as presented by Tian et
al [10]. Also, ammonia will decrease the reactivity of hydrogen due to its high concentration allowing a
better mixing due to higher residence times. This process will decrease emissions, limit temperatures and
stabilize the flame. This concords with other studies (to be published) where Lewis number varies as a
consequence of changes in mass-thermal diffusivity. Nevertheless, the experiments show the high 3-
Dimensionality of the flame, Fig 3, requiring phase locking at higher repetition rates in the near future.

Fig. 3. Instantaneous chemiluminescense comparison between pure methane and hydrogen/ammonia, respectively.

Numerical values showed a very good correlation to experimental measurements, Fig 4. Only CO, O2
and NOx are documented here, but values for other molecules such as H 2 , UHC, CO2 , NH3 , etc. were
also obtained. Of special interest are those regions after stoichiometry where best stability seems to be
obtained. The low amount of emissions, high temperature and good combustion profiles will allow the
development of an efficient combustion process close to this point. However, as previously seen in the
experiments, it is just a narrow range before the flame starts to vibrate through a thermoacoustic process,
meaning that care is required to avoid these unstable conditions.
 A. Valera-Medina et al. / Energy Procedia 75 (2015) 118 – 123 123

Fig. 4. Results using GASEQ (red symbol) and Emission analyzers (blue and green symbols). Very good correlation was obtained.

4. Conclusions
It has been shown that stable flames can be achieved with low emissions using strong swirling flows
with both CH4 -NH3 and H2 -NH3 blends. Good correlation exists with experiments and numerical
calculations. OH radical emissivity seems to be overwhelmed by other radicals, i.e. NH2 molecules that
keep reacting with NOx emissions. It is recognized that there is only a narrow equivalence ratio where
high stability, low emissions and high temperature can be achieved, leaving a vast field of research on
how to improve these parameters to obtain wider operational ranges.
Acknowledges
The authors gratefully acknowledge the support of Siemens AG through project 507698. The authors
would like to thank Mr. Steve Morris for his inputs, suggestions and hard work.
References
[1] Ahlgren W. Fuel Power Density, ASME J Pres Vessel Tech 2012; 134: 054504.
[2] Karabeyoglu A, Evans B, Stevens J, Cantwell B. Development of Ammonia Based Fuels for Environmentally Friendly
Power Generation, 10th Ann Int Energy Conv Eng Conf, IECEC 2012.
[3] Grcar JF, Glarborg P, Bell JB., Loren A, Jensen AD. Effects of Mixing on Ammonia Oxidation in Combustion Environment
Intermediate Temperatures, Int Symp Combust, Abstract, 2004; 87.
[4] Karabeyoglu A., Evans B. Fuel Conditioning System for Ammonia Fired Power Plants, NH3 Congress, Iowa, USA, 2012.
[5] Pratt DT. Report TS-67-5, 1967; pp.35.
[6] Verkamp FJ, Hardin MC, Williams JR. Ammonia Combustion Properties and Performance in Gas Turbine Burners. Int
Symp Combust 1967; 11(1): 985-992.
[7] Meyer T, Kumar P, Li M. Ammonia Combustion with Near-Zero Pollutant Emissions, NH3 Congress, Iowa, USA, 2011.
[8] C. Morely, 2010, GASEQ version 0.79. [online]. http://www.arcl02.dsl.pipex.com [Accessed 25th Sept 2014]
[9] Pearse RWB, Gaydon AG, The Identification of Molecular Spectra, 4th Edition, John Wiley & Sons, USA, 1976, p. 234.
[10] Tian Z, Li Y, Zhang L, Glarborg P, Qi F. Combust Flame 156 (2009) 1413-1426.

Biography
Dr. Agustin Valera-Medina is a Chartered Engineer, Lecturer at Cardiff University. He
works at the Gas Turbine Research Centre on topics such as combustion hydrodynamics,
fluid mechanics, alternative fuels and novel power generation cycles.